PROJECT SUMMARY Individuals? adaptive immune responses are central to the epidemiology and evolution of influenza and the effectiveness of influenza vaccines. It is therefore surprising that despite nearly 70 years of study, major questions about the immune response to influenza remain unanswered. In particular, it is unclear how well natural infection protects from reinfection with the same or related types and subtypes, how vaccination affects protection against symptomatic and asymptomatic infections over time, and how protection varies with immune history, age, individual, sex, and other factors. The two main obstacles to progress have been a shortage of observations from the same individuals over time and a lack of modeling approaches that can accommodate the complex, stochastic dynamics of infection and immune response replicated across individuals. The proposed research takes advantage of an extraordinary influenza cohort and new methods for longitudinal modeling to understand how protection to influenza infections of varying severity arises, and especially how it is shaped by infection and vaccination history. The ongoing Nicaragua Pediatric Influenza Cohort Study (NPICS) has followed thousands of children since 2011 and recorded their antibody titers, infections, symptoms, and vaccination history to influenza. We will use these data to fit and evaluate a large set of stochastic, individual-level, mechanistic, dynamical models to estimate the duration of protection and its dependence on exposure history and other factors. First, we will estimate the duration of protection against reinfection with the same type or subtype and evaluate its dependence on the order of early exposures and host and viral characteristics. Next, we will measure the strength and duration of cross-protection between type and subtypes. Finally, we will compare the dynamics of protection after natural infection to those after vaccination, including repeat vaccinations. Our flexible modeling approach takes advantage of diverse data types and inference techniques while allowing precise formulation of biological hypotheses mathematically. Its recent success with similar longitudinal datasets of PCR-confirmed viral infections and influenza serology demonstrates feasibility. Preliminary results suggest a role of exposure history on heterosubtypic infection risk. This work is poised to advance basic knowledge on influenza and the development of immune memory, and it will provide a new set of dynamical modeling tools for longitudinal data. This project will thus achieve NIH MIDAS objectives by advancing the development of inference techniques and software for an important and growing type of data and by expanding knowledge of an important host-pathogen dynamic. This work also directly addresses priorities established by the NIH Strategic Plan for the development of a universal influenza vaccine, especially identifying factors associated with the severity of influenza (objective 1.2) and improving understanding of how and when exposure to influenza antigens shapes the response to infection and vaccination (objective 2.1).